Use normalization instead of standardization?

[OleMussmann]

https://machinelearningmastery.com/normalize-standardize-time-series-data-python/

• Normalization (min-max scaling): values between 0 and 1
• Standardization (z-score normalization): mean is 0, standard deviation is 1

We currently use standardization, but intuitively normalization could be a more natural fit.

• [ ] Benchmark performance of both
• [ ] Change variable names in the code to statistically correct approaches

[sjvrijn]

Can be looked into once #64 is merged